This paper is about application of various data mining methods in social networks and social media area. It reveals basic principles of social media with the aim to high information potential of usage of the data from social networks. This is demonstrated on selected data mining methods, especially Social Network Analysis and Sentiment Analysis. Other opportunities of using social media data are shown in chapter about Social Media Monitoring tools. All these chapters are supplemented by practical examples and particular researches. Last chapter reveals visions and threats, which can bring data mining in the future. Keywords Data mining, social networks, social media, social network analysis, sentiment analysis, social media monitoring

This study was done to examine the relationship between self-rated health and social/community relations. Due to advances in modern medicine, multifactorial diseases are more prevalent than acute infectious diseases and a greater understanding of the impact sociological variables has on health is of great importance. In prior research, self-rated health has been demonstrated to be a robust predictor of mortality, even when controlling for other variables known to impact health. Presence of a strong social network and attachments to community have been shown to be protective of self-perceptions of health. The Health and Living study was conducted in the Bear River Health District located in northern Utah in 2004 utilizing a mail survey. The relationship between self-rated health and social network indicators in addition to community attachment variables was evaluated statistically. Demographic variables were also analyzed. Church attendance, number of friends, income, age, and education were found to be statistically significant.

The U.S. has undergone macroeconomic changes over the latter course of the twentieth century. As a result, migration patterns have shifted toward the fast-growing southern and western portions of the nation. My research measures the impact of deindustrialization and educational selection on out-migration from the metropolitan Rust Belt for 1980, 1990, and 2000. Analysis on destination selection using multinomial regression analysis is then conducted to determine whether education trumps social capital for long-distance migration. Findings indicate that more severely deindustrializing metropolitan areas have greater out-migration in 1980 and 1990 but less so for 2000, with positive educational selection for each year. Multinomial results indicate that education does not attenuate social capital for interregional migration destination. The rise of the service economy may indicate the increasing importance of social capital for individuals leaving the Rust Belt for other regions.

No description available.

No description available.

No description available.

This study investigated the social networks developed by language learners and their relationship with perceived changes in the learners' fluency in the context of an intensive English as a second language (ESL) program. Using data from the Study Abroad Social Interaction Questionnaire (SASIQ) to determine SN development and native speaker ratings to determine perceived fluency, a simple linear regression to test the relationship between social network variables and fluency gain, as well as a hierarchical regression measuring (a) the combined effect of variables previously found to be significant in fluency gain, and (b) the additional joint effect of the remaining social network variables. We found that participants were successful in developing complex social networks, and that their oral fluency did increase significantly in connection with their social networks. Density (the average number of people listed in a social group) was the most important factor when only social network variables were considered. In the hierarchical regression, initial proficiency level and the percentage of native English speakers in the network were the most significant of the established variables in the first step, and overall size and density were the most important of the added social network variables in the second step.

The ingrained economic, political, and cultural structures in society create difficulties for people to change consumption habits because they have existed in society for a long time where individuals’ social networks affect consumption. The purpose of this research has been to highlight and analyze the effect of the connection between social network and consumption and to reflect on how one's social network affects consumption habits. A quantitative survey using the 2018 Living Costs and Food Survey dataset with 5000 respondents was conducted for this reason. Social cost was used as a measure to highlight the effect it has on consumption. The result highlighted that material products acted as symbolic capital that the respondents consumed to strengthen their social ties that existed within their social networks.

No description available.

Analysis of large network datasets has become increasingly important. Algorithms have been designed to find many kinds of structure, with numerous applications across the social and biological sciences. However, a tradeoff is always present between accuracy and scalability; otherwise promising techniques can be computationally infeasible when applied to networks with huge numbers of nodes and edges. One way of extending the reach of network analysis is to sparsify the graph by retaining only a subset of its edges. The reduced network could prove much more tractable. For this thesis, I propose a new sparsification algorithm that preserves the properties of a random walk on the network. Specifically, the algorithm finds a subset of edges that best preserves the stationary distribution of a random walk by minimizing the Kullback-Leibler divergence between a walk on the original and sparsified graphs. A highly efficient greedy search strategy is developed to optimize this objective. Experimental results are presented that test the performance of the algorithm on the influence maximization task. These results demonstrate that sparsification allows near-optimal solutions to be found in a small fraction of the runtime that would required using the full network. Two cases are shown where sparsification allows an influence maximization algorithm to be applied to a dataset that previous work had considered intractable.